The present invention generally relates to apparatuses for reproducing signals which are pre-stored in a memory, and more particularly to an apparatus which reproduces a small level portion and a large level portion of an analog signal which is obtained by performing a digital-to-analog conversion, with a high resolution, when reproducing and subjecting to the digital-to-analog conversion a pulse code modulation (PCM) signal which is pre-stored in a memory.
In a conventional electronic organ of a first type, an output signal of an oscillator circuit is generally subjected to a waveform shaping, so as to obtain a signal waveform closely resembling the sound of a desired instrument. However, the sound of the instrument obtained by this electronic organ of the first type, is simply an artificially obtained sound which closely resembles the sound of the actual instrument. Accordingly, an electronic organ of a second type which is capable of obtaining a sound equivalent to the sound of the actual instrument, has been developed. In this electronic organ of the second type, the sound of the actual instrument is converted into a PCM signal and is stored in a digital memory. The sound is produced from the electronic organ by reading out the pre-stored PCM signal from the digital memory and then converting the read out signal into an analog signal. The electronic organ of the second type thus comprises a built-in memory which is prepared by the manufacturer, and the built-in memory pre-stores the PCM signal which is obtained by converting the sound of the actual instrument into the PCM signal. Hence, when the operator plays the electronic organ of the second type, the desired sound is read out from the built-in memory so as to produce a sound which is essentially the same as the sound of the actual instrument.
According to a conventional apparatus for storing the sound of the actual instrument in the memory, an audio signal corresponding to the sound of the actual instrument is sampled and is then subjected to a linear quantization so as to obtain a digital (PCM) signal. This digital signal is written into the memory. However, since the digital signal stored in the memory has a constant quantization width, the quantization noise becomes larger as the level of the input signal becomes smaller, and there is a problem in that the signal-to-noise (S/N) ratio becomes deteriorated.
Accordingly, there was an apparatus which samples the audio signal and then compares the sampled audio signal with a predetermined reference value. The level of the audio signal having a level smaller than the reference value is amplified, and the level of the audio signal having a level larger than the reference value is unchanged. The audio signal which is controlled of its level in this manner, is subjected to a quantization and is written into the memory. According to this conventional apparatus, the signal storage can be performed with a high resolution with respect to the input signal having a small level. Among a predetermined number of bits in the digital data which is stored and read out from the memory, at least one bit is used as a so-called gain bit for discriminating in a reproducing system whether the level is amplified. The remaining bits among the predetermined number of bits are used as a level code indicating the level information. For this reason, the number of bits of the level code, that is, the quantization number of bits in this conventional apparatus is smaller than the quantization number of bits in the conventional apparatus which performs the linear quantization, by the number of bits which are used for the gain bit. As a result, the resolution of the input signal having a large level becomes deteriorated. In addition, in a vicinity of a zero level (center level) of the signal waveform, the level is discriminated as being a small level and this zero level is amplified. Therefore, the level of the signal is controlled at various points on the signal waveform, and the original signal waveform cannot be reproduced accurately unless an attenuator having an accurate gain is used. Hence, there is a problem in that a crossover distortion occurs at the point where the level changes, and the points where the S/N ratio is poor becomes large.